ss Great Britain iron hull: modelling corrosion to define

storage relative humidity

David Watkinsona and Mark Lewis b
a

Senior Lecturer, Conservation Section, School of History and Archaeology, PO Box 909, Cardiff
University, Cardiff, CF10 3XU, UK.
b Research Assistant, Conservation Section, School of History and Archaeology, PO Box 909, Cardiff
University, Cardiff, CF10 3XU, UK.

Abstract
Brunels wrought iron ship ss Great Britain is corroding in its dry dock in Bristol. Corrosion
control will involve sealing and desiccating the dock. The corroding iron hull forms chloride
rich FeOOH and during desiccation ferrous chloride is likely to form. FeOOH/iron and
FeCl2 .4H 2 O/iron were subjected to controlled relative humidity (1%) and corrosion
thresholds were experimentally determined. Iron in contact with these compounds corroded
above the thresholds but not below them. Iron did not to corrode in the presence of
FeCl2 .2H 2 O. The stability of this compound to moisture was determined. Corrosion was
extremely slow just above corrosion thresholds, but increased significantly as relative
humidity values rose to 30%. These results are being used to set an operational relative
humidity for the desiccated environment around the ss Great Britain and to offer new
conservation options for other historic ships.
Keyword: iron, corrosion, ferrous chloride, -ferric oxyhydroxide, relative humidity, storage.

1. Introduction
1.1 ss Great Britain: Historical context
The 1843 wrought iron steamship ss Great Britain was salvaged from the Falkland Islands
in 1970 and placed in the Great Western Dockyard in Bristol, which had been constructed
especially for building the ship, making it the first integrated iron shipbuilding yard in the
world (figure 1). In its open air dry dock the iron hull of the ss Great Britain equilibrated with
its environment, which rarely drops below 70% due to climate, seepage of water through the
docksides, faulty drainage and a leaking caisson (Cox and Tanner 1999, Turner et al 1999).
Insert - Fig 1.
Initial evaporation of seawater from the hull left behind highly soluble and deliquescent
chloride salts, which are unlikely to be present in their solid crystalline form, due to the high
ambient relative humidity in the dock. They will be solvated within corrosion layers, micro
cracks, pits and in the overlaps between iron hull plates at rivet points. In this state their
contribution to the electrolytic corrosion of the iron will be significant; even though rain-wash
a

Corresponding Author: TEL: +44 (0) 29 20 874249: FAX: +44 (0) 29 20 874929: email:
Watkinson@cf.ac.uk

over the years and an aqueous pressure wash, which was used as an early conservation
method, will have reduced their concentration. The hull is rapidly corroding in the high
relative humidity within the dock environment and engineering reports estimate that it will be
structurally unsafe in about 25 years (Turner et al 1999).
1.2 Conservation programme
Early conservation and maintenance of the hull had involved 10,000 psi aqueous pressure
blasting and traditional shipyard practices that applied protective coatings (Turner et al
1999). Understandably these measures failed to either stop corrosion or to appreciably slow
its progress (Figure 2).
Insert - Fig 2
A long-term development plan adopted a mission statement in 1998:
To preserve the ship, ss Great Britain, and its buildin g dock for all time for the public
benefit of all, and to place the same upon public display as a museum for the enhancement of
public understanding and appreciation of her social, commercial, scientific and technological
context and significance.
The decision from the ss Great Britain Trust to conserve fully to the highest standards
the extant fabric of the ss Great Britain (1843-1970) and its building dock (Cox and Tanner
1999) defined future preservation by firmly recognising it was a conservation task, rather than
a restoration programme. The corroded wrought iron plates of the hull are considered an
essential part of the ships fabric and are to be conserved in situ, with their corrosion in place,
rather than replaced with mild steel replicas.
The current conservation plan identified the condition of the ship and reviewed
conservation options (Turner et al 1999). It was clear that the condition of the iron and the
length (322 feet) of the hull rules out many conservation approaches. Methods employing
stripping or traditional surface treatments for ships would damage the badly corroded hull
(Table 1). Treatments designed to remove or inhibit the action of chlorides are unpredictable
and offer technical limitations on a large scale project (Table 2). The amount of chloride that
remains in iron after chloride extraction procedures have been applied cannot be known. It
will remain as an unknown threat to the future stability of the iron.
Insert - Table 1
Insert - Table 2
In order to employ a conservation method whose effectiveness could be successfully
modelled and tested prior to implementation, it was decided to conserve the iron hull by
controlling the availability of water to support the corrosion processes (Turner et al 1999).
Desiccation is the both the least interventive and least unpredictable method for preserving
the ss Great Britain. As part of their ethical framework the ss Great Britain Trust was pleased
to be able to choose a treatment option that retained the iron corrosion layers on the hull.
To implement this conservation option the corrosion processes occurring on the hull need
to be identified and their response to relative humidity assessed by modelling corrosion in the
laboratory. In this way the degree of desiccation required to prevent them occurring can be
measured. This data can then be used to design a controlled storage environment that
moderates or prevents corrosion.

Desiccating the environment around the hull is a technically challenging task that
involves constructing an envelope around the ship and controlling the relative humidity
within this space. Consulting with architects and climate control engineers produced an option
that would roof the dry dock with glass, from the waterline of the ship to ground level at the
dockside. The interior of the ship would be modified to produce an environment with
minimum air leakage between the lower and upper decks. Within this space encompassing the
exterior and interior of the hull below its original water line, moisture levels will be controlled
to a value that will either dramatically reduce or prevent chloride and chloride bearing
compounds contributing to corrosion of the iron hull. Visitors would have access to both the
interior of the hull below the waterline and the dry dock facility (Figure 3). The hull will
appear to float in its dry dock, as the roof will have water across it. Above the waterline the
hull has lower chloride levels and is in good condition, making it suitable for a coating
system.
Insert - Fig 3
1.2 Desiccation for corrosion prevention
Desiccation of iron has been employed for over 40 years, with varying degrees of success,
to store and preserve archaeological iron (Watkinson and Neal 1998). Remarkably, very little
investigation of this process has been carried out, yet it remains the storage mainstay for
archaeological iron. Turgoose (1982a and b) and Selwyn et al. (1999). provided some
information on the corrosion of chloride infested iron during burial and following excavation,
there are no accurate figures on the level of desiccation necessary to prevent corrosion of
chloride infested iron.
The design for successful desiccation depended on the ss Great Britain Project identifying
how relative humidity affected corrosion reactions that could occur on the hull. The cost of
dehumidifying an environment is closely linked to the target relative humidity. Very low
relative humidity values will cost much more money to achieve and sustain, due to plant
requirements for dehumidification and tolerances in the architectural design of the controlled
space.
The ss Great Britain Trust sought to:

Determine how corrosion reactions occurring on the iron hull were influenced by
relative humidity.
Identify the relative humidity at which corrosion of iron in chloride infested
environments ceases.
Explore how the rate of corrosion of chloride infested iron changes with relative
humidity.
Use this data to model a storage environment that will enhance the life of the iron hull
by either preventing or reducing corrosion rate according to the resources available to
them;
Offer this model as part of an 8.4 million bid to the Heritage Lottery Fund for
support to conserve the ss Great Britain and its historic dockyard.

1.3 Corrosion of the ss Great Britain iron hull

Analysis of the hull revealed significant amounts of soluble chloride in the corrosion
product and the presence of FeOOH (Akageneite), Fe 3 O4 (Magnetite) and other FeOOH
(ferric Oxy-hydroxide) polymorphs. FeOOH was identified by XRD (Figure 4) and Soxhlet
extraction of corrosion products from various parts of the hull revealed chloride levels within
the 160 to 3200 ppm range (Turner et al 1999). Since chloride is concentrated at anode sites
at the metal surface overall chloride levels in the hull are likely to be much higher than the

measurements suggest (Ishikawa et al 1988, Selwyn et al 1999, Turgoose 1982a,Watkinson

1983).
Insert - Fig 4
The presence of these corrosion products fits the corrosion model suggested by Turgoose
(1982a and 1982b) in which Fe 2+ is oxidised and resulting acid can attack iron directly to free
more Fe 2+. There is also free chloride in solution acting as an electrolyte (Equation 1). Selwyn
et al (1999) cited a cycle involving acid attack on iron by HCl, which is also described for
pitting by Jones (1992).
4Fe2+

O2 + 6H 2O

4FeOOH +

8H +

[1]

In the environment within in the dock area the iron hull is currently corroding rapidly due
to high ambient relative humidity that allows the soluble chloride to act as an electrolyte to
support corrosion. This situation will change when the hull is dried within its new controlled
storage environment. Chloride will concentrate as the moisture within the iron evaporates.
Anode sites will become increasingly acid due to hydrolysed Fe 2+ and this provides low pH
conditions in which solid FeCl2 .4H 2 O may form (Jones 1992, Turgoose 1982a).
1.4 Low humidity corrosion of chloride infested iron
Turgoose (1982a) identified iron corrosion processes involving FeCl2 .4H 2 O and FeOOH
that occur at low relative humidity. Iron corrodes when mixed with ferrous chloride at 20%
relative humidity, but it does not corrode when the relative humidity is 15% (Turgoose
1982b). This is attributed to the stability of FeCl2 .4H 2 O at the higher humidity and
FeCl2 .2H 2 O at the lower humidity. The water in FeCl2 .4H 2 O is thought to support electrolytic
corrosion, whereas FeCl2 .2H 2 O does not contain enough water to do this.
Turgoose (1982b) postulated that FeOOH is hygroscopic and that this contributes to
corrosion of iron in contact with it. FeOOH contains chloride in its crystal structure and
adsorbs it onto its surface when it forms in chloride rich environments (Ishikawa et al 1988).
This chloride can become mobile in moisture that is attracted to its surface and it may be this
that facilitates either the electrolytic corrosion of iron or direct acid attack on iron (Lewis and
Watkinson, forthcoming).
1.5 FeOOH on the hull of the ss Great Britain
The hull of the ss Great Britain will carry two forms of FeOOH. The first will have
formed in the atmosphere during the past 35 years. Much of this FeOOH is subject to rain
wash, which is likely to have removed most of its surface adsorbed chloride. If mobile surface
adsorbed chloride causes iron in contact with FeOOH to corrode, then washed FeOOH
should not corrode iron.
FeOOH will also be present as a corrosion product in areas of the ship that are not rainwashed and it will form as iron corrodes during the proposed desiccation of the hull. This
FeOOH will retain its surface adsorbed chloride and present a corrosion threat to iron that it
is in contact with, provided the storage relative humidity supports the FeOOH/iron corrosion
reaction.
It is reported that FeOOH is metastable with respect to aFeOOH over a 25 year time
period (Gilberg and Seeley 1981). This offers dangers for the ship as its transformation to
FeOOH will release chloride that can potentially act as an electrolyte. An FTIR assay of a
23-year-old sample of FeOOH that was stored in a closed glass jar in the conservation
laboratories at Cardiff University, revealed the same composition as when it was produced. It
is also known that pH influences transformations of FeOOH. These are areas that require
further study.

Based on the corrosion model involving FeOOH and ferrous chloride, experiments were
carried out to link relative humidity to corrosion of chloride infested iron in low humidity
environments. Results were used by the ss Great Britain Trust to design their storage
environment.
2. Experimental Procedure
2.1 The experiments carried out in this study aimed to:
Confirm that iron does not corrode in the presence of FeCl2 .2H 2O
Identify the highest relative humidity at which FeCl2 .2H 2 O exists before it converts to
FeCl2 .4H2 O
Examine the corrosion of iron mixed with FeCl2 .4H 2 O at various relative humidities.
Establish the relative humidity at which iron mixed with FeOOH ceases to corrode.
Examine the corrosion of iron mixed with unwashed FeOOH at selected relative
humidities.
Remove surface adsorbed chloride from FeOOH (this simulates rain-wash of this
corrosion product) and determine whether washed FeOOH corrodes iron.
Results of the experiments were used to recommend:
A relative humidity below which iron does not corrode if mixed with ferrous chloride
A relative humidity below which iron does not corrode if mixed with FeOOH
Offer comment on the rate of corrosion of ferrous chloride/iron and FeOOH/iron
mixtures at relative humidity values that were slightly above the no-corrosion
relative humidity identified above.
This data was used by engineers and architects to design climate control plant and the
tolerances of the dock and hull seals that would create the storage space around the ss Great
Britain Hull.
2.2 Research method
Analar grade FeCl2 .4H 2O and pure iron powder were purchased and FeOOH was
synthesised by exposing pure iron powder and ferrous chloride to 92% relative humidity until
the iron oxidised completely, which took several months. The resulting assay showed 99%
FeOOH and 1% FeOOH (Figure 4).
Washed FeOOH was produced by subjecting synthesised FeOOH to aqueous Soxhlet
wash until no more chloride was being released into the wash solution, as determined by
specific ion meter. At this point it was deemed that mobile surface chloride is lost from the
FeOOH and only the chloride locked in the crystalline hollandite tunnel structure of the
FeOOH would remain (Childs et al 1980, Gilberg and Seeley 1981, Ishikawa and Inouye
1972, Ishikawa and Inouye 1975). Unless otherwise stated all FeOOH used in experiments
reported here is unwashed.
Testing involved the use of a climatic chamber that controlled relative humidity to 1%
(established by test calibrations) and +0.5o C. Test samples were placed on a balance
(accuracy 0.0001g) in the chamber. The balance was calibrated for accuracy and drift using a
standard weight, exposed to 20% relative humidity at 20o C for several days. Systematic error
was reproducible and slight balance drift could be accounted for when analysing trends.
The chamber can be programmed to operate either at a particular relative humidity or on a
particular relative humidity programme. Weight changes were dynamically monitored to file
every 5 minutes. Visual inspection, XRD and FT-IR were used to determine change in the
sample composition and attribute weight change to one or a combination of; iron corrosion;
oxidation of corrosion product; desiccation or hydration. In this way the influence of relative

humidity on the corrosion process could be monitored and graphic results show weight
variation of test samples against time for fixed humidity values. All tests were standardised at
20o C enabling conversion between relative and specific humidity. For consistency results are
hereafter reported in terms of relative humidity.
3. Results
Samples of FeCl2 .2H 2O and FeCl2 .4H 2 O were mixed with iron powder and exposed to
differing relative humidities. A sample comprising FeCl2 .2H 2 O and iron powder did not gain
weight after 20000 minutes (13.9 days) at 19% relative humidity (Figure 5). This showed that
FeCl2 .2H 2 O is the stable form of ferrous chloride at 19% relative humidity and iron does not
corrode in the presence of FeCl2 .2H 2O. Both hydration to FeCl2 .4H 2 O and oxidation of iron to
FeOOH would have produced a weight increase. In contrast FeCl2 .4H 2O mixed with iron
powder gained weight at 22% and 27.5% and 35% relative humidity. This confirmed that
FeCl2 .4H 2 O is the stable form of ferrous chloride at 23% relative humidity and above and that
it corrodes iron in contact with it. Corrosion is slow at 22% relative humidity but speeds up
appreciably at 35% relative humidity.
Insert - Fig 5
The corrosive effect of FeCl2 .4H 2O on iron at different relative humidities is seen in
figure 6. Samples comprising 2g of FeCl2 .4H 2 O mixed with 2g of iron powder were exposed
to differing relative humidities. Percentage weight change at 4000 minutes is compared.
Below 25% relative humidity corrosion is limited. As relative humidity rises corrosion
increases and is ten times greater at 30% relative humidity as compared to the rate at 25%
relative humidity.
Insert - Fig 6.
Using the 19% no-corrosion marker for iron chloride mixed with iron powder determined
in figure 6, a mixture of FeOOH and iron powder was exposed to 19% relative humidity for
4000 minutes (2.8 days). Initially weight was lost due to dehydration of the FeOOH and this
masked any corrosion taking place. At 2000 minutes the sample started to gain weight due to
corrosion of the iron, whose contribution to weight gain exceeded any continued weight loss
due to dehydration (Figure 7). Corrosion then continued at a slow steady rate. Iron in contact
with unwashed FeOOH corrodes at 19% relative humidity.
Insert - Fig 7
Mixtures of unwashed FeOOH and iron powder were exposed to a range of relative
humidities to determine the point at which the corrosion of iron ceased when in contact with
unwashed FeOOH (Figure 8). All samples showed a rapid initial desiccation of FeOOH.
Iron/FeOOH mixtures did not measurably corrode at 12% relative humidity over the
duration (16000 minutes) of the test but clearly corroded at 15% relative humidity.
Insert - Fig 8.
A sample of washed FeOOH/iron powder was exposed to 19% relative humidity. When
compared to the unwashed sample at 19% relative humidity (figure 7), it is clear that removal
of surface chloride has eliminated the immediate aggressive corrosive effects of the FeOOH
on iron (figure 9). The lack of weight loss by the sample may indicate that surface adsorbed

chloride is responsible for the majority of the hygroscopic behaviour of FeOOH. Our initial
comments are that this may be due to the association of chloride with water and subsequent
production of HCl on the surface of the FeOOH, which then attracts water. Once chloride
has been removed surface sites are occupied by hydroxyl ions from the wash solution. More
research is required to investigate this theory.
Insert - Fig 9.
A mixture of 2g quantities of FeCl2 .4H 2 O, unwashed FeOOH and iron powder was
exposed to 20% relative humidity. A significant loss in weight occurred from dehydration of
the corrosion products during the first 12000 minutes (8.3 days) (figure 10). This was
followed by steady weight gain due to corrosion of the iron. Corrosion of iron in contact with
FeOOH will occur at 20% relative humidity. Since FeCl2 .2H 2 O is stable at 19% relative
humidity and FeCl2 .4H 2 O at 22% relative humidity, any hydration of FeCl2 .2H 2 O will be
slow at 20% relative humidity. Any contribution from ferrous chloride to corrosion will be
negligible in the time frame of the experiment here.
Insert - Fig 10.
4. Discussion
The results reported here endorse the principle of desiccation as a means to preserve the
iron hull of the ss Great Britain. The hull of the ss Great Britain contains a range of iron
corrosion products that include FeOOH. As it dries in its new storage environment, it is
likely that FeCl2 .4H 2 O and fresh (unwashed) FeOOH will form. The experiments reported
here show that FeOOH and FeCl2 .4H 2 O are capable of corroding iron that is in contact with
them. To prevent FeCl2 .4H 2O corroding iron relative humidity will have to be lowered to 19%
to form FeCl2 .2H 2 O. To prevent FeOOH corroding iron requires a relative humidity of 12%
or less. Where both compounds are present a relative humidity of 12% or less is necessary to
stop the corrosion of iron in contact with them. If FeOOH has been washed to remove
surface adsorbed chloride, it does not measurably corrode iron in comparison to unwashed
FeOOH. Rain-wash of the ships hull may have produced this effect on FeOOH on its
surface.
Desiccating the 322 feet long hull of the ss Great Britain to 12% relative humidity would
be costly and technically challenging, with potentially high maintenance costs. Examining the
influence of FeCl2 .4H 2 O and FeOOH on the rate of iron corrosion, as measured by the
weight increases shown in figures 6 - 8 indicates that corrosion is many times slower at 20%
relative humidity, as compared to 25% or 30% relative humidity. It is worth taking a
pragmatic view of the FeOOH /FeCl2 .4H 2 O iron corrosion model and considering the
overall effect of relative humidity on the rate of corrosion (figure 11). The stability of the iron
hull of the ss Great Britain could be significantly enhanced and its life-span prolonged by
storage environments that control relative humidity to 20%.
Insert - Fig 11.
Cost effectiveness often promotes compromises. The results of these experiments allow
informed decisions as to the effectiveness of a particular humidity for the preservation of
chloride infested iron. Capital expenditure and future income can be related to best value
decisions for the long-term preservation of the iron hull (Table 3). At the ss Great Britain it is
planned that costs will be minimised by controlling relative humidity in the immediate
vicinity of the hull. The ss Great Britain Trust plans to force dried air over the hull to create
and sustain a localised relative humidity that prevents or significantly reduces corrosion of the
type examined in this research. Currently experiments are underway to determine the effects
of small short-lived fluctuations in relative humidity above the chosen control relative

humidity. Simulations of the effect of catastrophic plant failure on corrosion rates are also
underway. The results of this work successfully supported a 8.4 million grant from The
Heritage Lottery Fund in the UK.
Insert Table 3
The results of these laboratory experiments are discussed in the context of the ss Great
Britain and the juxtaposition of these corrosive products and iron on the hull must be
considered in any corrosion model. The large surface area contact provided by powders used
in the experiments and the large quantities of reactants are not duplicated on the hull. The
physical nature of the corrosion layers on the ship will be likely to have a very significant
effect on reaction rates. However, experience with chloride infested archaeological iron
suggests that corrosion can be rapid and invariably occurs at the interface between metal and
corrosion layers, creating spalling of the corrosion layers.
The results of this research and their successful implementation offer a new model for the
conservation of large iron ships. Both the level of relative humidity and its effects have now
been linked to specific chloride based corrosion reactions. When compared to chloride
removal treatments for corroded iron, desiccation can now offer greater predictability, with
the option of monitoring effectiveness to maintain standards. It also provides its user with
measurable choices. Since its efficacy can be graded according to the prevailing relative
humidity, it is possible to model the level of corrosion control that a particular sum of money
can buy, in terms of plant and operational costs. The application of this conservation
technique can be extended to other chloride contaminated iron objects, both large and small.
Acknowledgements
The authors thank the ss Great Britain Trust, Heritage Lottery Fund and the Director of the
ship, Matthew Tanner, for sponsoring this research. Our thanks also to Robert Turner of Eura
Conservation Ltd. For access to his reports and useful discussion. XRD was kindly carried out
by Dr Anthony Oldroyd, School of Earth Sciences, Cardiff University.
References
Argryropoulos, V., Selwyn, L. S. and Logan, J. A. (1997). Developing a conservation
treatment using ethelyenediamine as a corrosion inhibitor for wrought iron objects found at
terrestrial archaeological sites. In Macleod, I.D., Pennec, S. and Robbiola, L.(eds.) Metal 95
Proceedings of Metals in Conservation Conference 25-28 September 1995, 153-8 James and
James, London
Barker, B.D., Kendell, K. and OShea, C. (1982). The Hydrogen Reduction Process for the
Conservation of Ferrous Objects. In. Clarke, R. and Blackshaw, S. (eds.) Conservation of
Iron National Maritime Museum Monograph, 53, 23-27 National Maritime Museum,
London.
Barker D., Johnston A. and OShea C. (1997). Conservation of the submarine Holland: a
practical overview. In Macleod, I.D., Pennec, S. and Robbiola, L.(eds.) Metal 95 Proceedings
of Metals in Conservation Conference 25-28 September 1995, 286-90 James and James,
London
Birchenall, E.C. and Meussner, R.A. (1977). Principles of Gaseous Reduction of Corrosion
Products. In Brown, F. B.(ed.) Corrosion and Metal Artifacts: A dialogue between
Conservators, and archaeologists and Corrosion Scientists. National Bureau of Standards
Special publication 479, 39-58.

Childs, C. W., Goodman, B. A., Paterson, E. and Woodhams, F. W. D. (1980). The Nature of
Iron in Akaganeite (-FeOOH). In Australian Journal of Chemistry 33, 15-26.
Cox, J. with Tanner, M. (1999). Conservation plan for the Great Western Steamship
Company Dockyard and the ss Great Britain. Volume 1. ss Great Britain Trust.
Gilberg, M.R. and Seeley, N.J. (1982). The Alkaline Sulphite reduction process for
archaeological iron: a closer look. Studies in Conservation 27, l80-4.
Gilberg, M.R. and Seeley, N.J. (1981). The identity of compounds containing chloride ions in
marine iron corrosion products: a critical review. Studies in Conservation 26, 50-6.
Ishikawa, T. and Inouye, K. (1972). The Structural Transformation of Ferric Oxyhdroxides
and Their Activity to Sulfur Dioxide. Bulletin of the chemical Society of Japan, 45, 23502354.
Ishikawa, T. Sakaiya, H. and Kondo, S. (1988). Adsorption of water on colloidal iron(III)
oxide hydroxides by infrared spectroscopy. Journal Chemical Society Faraday Transactions
I, 84, 1941-8
Ishikawa, T. and Inouye, K. (1975). Role of Chlorine in -FeOOH on Its Thermal Change
and Reactivity to Sulfur Dioxide. Bulletin of The Chemical Society of Japan, 48 (5), 15801584.
Jones, D. (1992). Principles and Prevention of Corrosion.. Macmillan. New York.
Knight, B. (1997) The stabilisation of archaeological iron : past present and future. In
Macleod, I.D., Pennec, S. and Robbiola, L.(eds.) Metal 95 Proceedings of Metals in
Conservation Conference 25-28 September 1995, 36-42, James and James London
North, N.A. and Pearson, C. (1975). Alkaline sulphite reduction treatment of marine iron.
ICOM Triennial Venice, 1-13 (75/l3/3)
North, N.A. and Pearson, C. (1978) Washing methods for chloride removal from marine iron
artifacts Studies in Conservation 23, l74-86.
Rinuy, A. and Schweizer, F. (1982). Application of the Alkaline Sulphite Treatment to
Archaeological Iron: A Comparative Study of Different Desalination Methods of
Conservation of Iron. In Clarke, R. and Bradshaw, S. (eds.) Conservation of Iron. National
Maritime Museum Monographs and Reports, no. 53, 44-49. National Maritime Museum
London
Scott, D.A. and Seeley, N.I. (1987). The washing of fragile iron artifacts. Studies in
conservation 32, 73-76.
Selwyn, L.S. and Logan, J.A. (1993). Stability of treated iron: A comparison of treatment
methods. In Bridgland, J. (ed). International Council for Museums Committee for
Conservation 10 th Triennial Meeting Washington DC. 803-07. James and James
Selwyn, L.S., McKinnon, M.R. and Argyropoulos, V. (2001). Models for chloride ion
diffusion in archaeological iron. Studies in Conservation 46, 109-121
Selwyn, L. S., Sirois, P.J. and Argyropoulous , V. (1999). The corrosion of archaeological iron
with details on weeping and akageneite. Studies in Conservation 44, 217-232

Skerry, B. (1985). How corrosion inhibitors work. In Keene, S.V. (ed.) Corrosion Inhibitors
in Conservation. UKIC occasional Paper 4., 5-12, UKIC London
Turgoose, S. (1982a). The Nature of Surviving Iron Objects. In Clarke, R. and Blackshaw, S.
(eds.), Conservation of Iron National Maritime Museum, Monograph No. 53, 1-8, National
Maritime Museum, London
Turgoose, S. (1982b). Post excavation changes in iron antiquities. Studies in Conservation 27,
97-101.
Turgoose, S. (1985) The corrosion of Archaeological Iron During Burial and Treatment.
Studies in Conservation 30, 13-18.
Turner, R., Tanner, M. and Casey, S. (1999). Conservation plan for the Great Western
Steamship Company Dockyard and the ss Great Britain. Volume 2 Condition report and
recommendations for the ss Great Britain. ss Great Britain Trust.
Tylecote, R.F. and Black, J.W.B. (1980). The effect of hydrogen reduction on the properties
of ferrous materials Studies in Conservation 25, 87-96.
Watkinson, D. and Neal, V. (1998). First Aid for Finds. UKIC.
Watkinson, D.E. (1996) Chloride extraction from archaeological iron : comparative treatment
efficiencies. In Roy, A. and Smith, P (eds.), Archaeological Conservation and Its
Consequences. International Institute for Conservation Copenhagen Congress 208-212. IIC
London
Watkinson, D.E. (1982) An Assessment of Lithium Hydroxide and Sodium Hydroxide
treatments for archaeological ironwork. In Clarke, R. and Blackshaw, S. (eds.), Conservation
of Iron National Maritime Museum, Monograph No. 53, 28-40. National Maritime Museum
London
Watkinson, D.E. (1983) Degree of mineralisation: its significance for the stability and
treatment of excavated ironwork. Studies in Conservation 28, 95-90.

10

Condition
Good condition
Areas covered with GRP, concrete or timber
Areas severely corroded likely to disappear upon abrasive cleaning
Areas in fair condition much iron likely to survive abrasive cleaning
Areas of severe corrosion likely to be perforated by abrasive cleaning

Percentage of
hull
3%
8%
17%
29%
43%

Table 1.

Treatment
Aqueous wash
Alkaline sulphite

Sodium hydroxide
Alkali wash
Electrolysis

Hydrogen reduction
Inhibitors
Cathodic protection

Comment on treatment within context of ss Great Britain conservation plan

Inefficient and unpredictable level of chloride extraction; aerated environments
provide lower chloride extraction than deaerated (Selwyn and Logan 1993, Scott
and Seeley 1987, Watkinson 1982 and 1996).
Better aqueous chloride extraction due to alkali and deaerated environment, but
normally less than 90% chloride removed; caustic solution; impractical on scale
required (Gilberg and Seeley 1982, Knight 1997, North and Pearson 1975, Rinuy
and Schweizer 1982, Selwyn and Logan 1993,Turgoose 1985, Watkinson 1996).
Between 55% and 80% chloride removed in aerated environments; caustic
solution; impractical scale (North and Pearson 1978, Watkinson 1982, 1983
and1996).
Sodium Carbonate wash; used with Holland 1 submarine; efficiency as chloride
remover unknown; practical application problems with very corroded structures;
disposal of treatment solution (Barker et al 1997, Knight 1997).
Varying reports on efficiency and stability; disposal of electrolyte is a problem;
challenging on a large scale; better with substantial metal cores; hull is
extensively mineralised in places (Knight 1997, Selwyn and Logan 1993,
Selwyn et al 2001).
Effective chloride removal; reactive pyrophoric iron produced; ethically
questionable; impractical on scale required (Barker et al 1982, Birchenall and
Meussner 1977, Tylecote and Black 1980).
Inhibitor action is unpredictable on heavily corroded iron surfaces that are
contaminated with chloride; toxicity limits choice (Argryropoulos et al 1997,
Skerry 1985, Turgoose 1985) .
Hull lacks metallic continuity due to areas of total mineralisation that are being
retained as part of the conservation rationale.

Table 2

11

Storage
category
very poor

enhanced
optimal

Relevance for the ss Great Britain

Current storage conditions for the ss Great Britain. Humidity lies
above the deliquescence point of ferrous chloride (~55% relative
humidity). Corrosion will be accelerated very greatly due to the
presence of the very strongly conducting electrolyte.
Humidity is below the deliquescence point for iron chloride and above
the corrosion threshold established for ferrous chloride/iron mixtures
in this research. Within this region the rate of corrosion increased
significantly with increasing relative humidity.
Any relative humidity beneath the threshold humidity for the
corrosion of ferrous chloride and iron powder mixtures. Ferrous
chloride no longer contributes directly to the corrosion of the iron.
FeOOH still contributes to corrosion.

Table 3

12

Captions
Table 1. Assessment of condition of the ss Great Britain hull (data from 2 )
Table 3. Storage categories for chloride contaminated wrought iron.
Table 2. Limitations of more traditional interventive methods used on chloride infested iron.

13

Figures

Figure 1.

Figure 2.

14

Figure 3.

Figure 4.

15

Figure 5.

Figure 6.

16

Figure 7.

Figure 8.

17

Figure 9.

Figure 10.

18

VERY
HIGH

Rate of Corrosion

HIGH

Aqueous solution
of electrolyte

VERY
POOR
STORAGE

LOW
Deliquescence Pt

ENHANCED
STORAGE

VERY
LOW

Threshold
R.H.
OPTIMAL STORAGE

Ferrous Chloride
4-hydrate stable

2-hydrate stable

0
0

20

50

Relative Humidity
(%)

Figure 11.

19

70

Figure Captions
Figure 1. ss Great Britain in the historical Great Western Dockyard Bristol.
Figure 2. Ongoing corrosion of the ss Great Britain hull in dry dock.
Figure 3. Artists impressio n of the controlled dock space enclosing the ss Great Britains
hull.
Figure 4. Comparison of XRD spectra for laboratory produced FeOOH (Akaganeite)
(bottom) and a sample taken from the hull of the ss Great Britain (top). The broad hump in
the upper spectra was due to mounting the sample on a glass plate. (International powder
diffraction files JCPDS 42-1315 and 34-1266.)
Figure 5. The line at 19% relative humidity is FeCl2 .2H 2 O mixed with iron powder. The lines
at 22%, 27.5% and 35% relative humidity are FeCl2 .4H 2O mixed with ir on powder. All
samples exposed for 14 days
Figure 6. The response of 2g FeCl2 .4H 2O /2g iron powder mixtures to relative humidity.
Increase in mass as a percentage, after 4000 minutes at a particular humidity.
Figure 7. FeOOH mixed with iron powder exposed to 19% relative humidity for 4000
minutes.
Figure 8. Samples comprising 2g of FeOOH mixed with 2g of iron powder exposed to 12%,
15%, 16% 21% and 25% relative humidity.
Figure 9. Washed FeOOH/iron powder mixture exposed to 19% relative humidity, as
compared to an unwashed sample of FeOOH/iron powder.
Figure 10. Unwashed FeOOH/FeCl2 .4H 2O/iron powder mixture exposed to 20% relative
humidity.
Figure 11. Schematic diagram of proposed corrosion rate/ relative humidity model.

20